Improving the performance in crop water deficit diagnosis with canopy temperature spatial distribution information measured by thermal imaging

Abstract

Infrared thermal imaging cameras are effective tools to monitor the spatial distribution of canopy temperature (Tc), which is the basis of the crop water stress index (CWSI) calculation. This paper presents a new method to improve the CWSI performance in crop water stress diagnosis based on Tc measured by thermal imaging. Cumulative frequency curves of pixel Tc extracted from each thermal image were analysed. Different statistical quantiles of Tc were determined, and the average Tc over different statistics quantiles were used to calculate the CWSI separately. There were large gaps among the CWSI based on Tc over different statistical quantiles. We compared the coefficient of determination (R2) of relationships among the CWSI based on Tc over different statistical quantiles and relative leaf photosynthetic activities. The general sensitive CWSI showed the best correlations with leaf photosynthetic activities, which were calculated based on average values of the top 60%, 50%, 70%, 50% of Tc statistics at different growth stages. The ranges of the CWSI with optimal leaf water use efficiency (between turning-points for downward trends in photosynthesis and transpiration) were 0.556–0.569, 0.481–0.486, 0.571–0.641, and 0.511–0.606 at tillering, panicle initiation to booting, heading to anthesis, and milk to soft dough stages respectively. The corresponding soil moisture levels were consistent with the lower thresholds of the rice under control irrigation. Based on the spatial distribution of canopy temperatures measured by thermal imaging cameras, the general sensitive CWSI, which was calculated by removing low temperatures, had a better performance in crop water stress diagnosis.